extracting stabilized image regions from handheld image sequences. dirty hack all around. consider this more of a playground repo that I can use as a base when quickly trying something.
python -m venv venv
. venv/bin/activate
pip install -r requirements.txt
python align.py input-path output-path